1 Executive Summary

1.1 Overview

This study examines transport services and mobility challenges in Kumasi Metropolitan Area, Ghana, based on surveys of 393 passengers and 339 transport operators across four major routes.

1.2 Key Findings

2 Sample Characteristics

2.1 Passenger Demographics

Table 1: Passenger Demographic Characteristics (N=393)
Characteristic Value
Sample Size 393
Female (%) 53.4%
Male (%) 46.6%
Mean Age Group 26–35 years (modal)
Tertiary Education (%) 23.7%
Self-employed/Trader (%) 30.3%
Income < GHC 1,000 (%) 19.6%
Long-term Residents (>7 years) (%) 69.2%

2.2 Detailed Demographic Breakdown

Table 2: Detailed Demographic Distribution
Group Category Count Percentage
Gender
Gender Female 210 53.4%
Gender Male 183 46.6%
Age Group
Age Group 18–25 years 106 27.0%
Age Group 26–35 years 100 25.4%
Age Group 36–45 years 87 22.1%
Age Group 46–55 years 59 15.0%
Age Group 56 years and above 27 6.9%
Age Group Below 18 years 14 3.6%
Education
Education Junior High School (JHS) 92 23.4%
Education No formal education 21 5.3%
Education Primary education 35 8.9%
Education Senior High School (SHS) 152 38.7%
Education Tertiary (College/University/Polytechnic) 93 23.7%
Income
Monthly Income (GHS) Below GHC 500 37 10.1%
Monthly Income (GHS) GHC 1,000 – 1,499 66 18.1%
Monthly Income (GHS) GHC 1,500 – 1,999 88 24.1%
Monthly Income (GHS) GHC 2,000 and above 134 36.7%
Monthly Income (GHS) GHC 500 – 999 40 11.0%

2.3 Driver Demographics

Table 3: Transport Operator Characteristics (N=339)
Characteristic Value
Sample Size 339.0
All Male (%) 100.0
Age 26-45 years (%) 54.3
Basic Education (%) 36.3
Trotro Drivers (%) 54.3
>5 Years Experience (%) 49.6
Union Members (%) 66.4
Work Full Day (%) 69.9

3 OBJECTIVE 1: Range of Transport Services

To identify the range of transport services provided in the Kumasi Metropolitan Area

3.1 Transport Mode Distribution

Table 4: Primary Transport Mode Usage (N=393)
Transport Mode Users Percentage % of Sample
Trotro (Minibus) 367 93.4 93.4%
Taxi 114 29.0 29%
Tricycle (Pragya) 73 18.6 18.6%
Motorcycle (Okada) 19 4.8 4.8%

3.1.1 Visualization: Transport Modes

3.2 Route-Specific Analysis

Table 5: Transport Service Distribution by Route
study_route Sample (n) % Trotro % Taxi % Tricycle % Motorcycle Avg Fare (GHS) Modal Diversity
Ejisu– Kejetia Road 123 90.2 17.9 7.3 0.8 6.6 2
Sofoline – Adum Road 107 98.1 55.1 15.0 5.6 15.6 3
Adum- Tafo Road 87 100.0 16.1 27.6 10.3 7.9 4
Kejetia– Santasi Road 76 84.2 25.0 31.6 3.9 8.8 3

3.2.1 Visualization: Routes and Modes

3.3 Travel Purpose Analysis

Table 6: Primary Travel Purposes
Purpose Count Percentage % of Trips
Work/Business 295 75.1 75.1%
Shopping/Errands 155 39.4 39.4%
Social/Religious 111 28.2 28.2%
School/Education 63 16.0 16%

3.4 Transport Frequency

Table 7: Weekly Public Transport Usage Frequency
Weekly Frequency Count Percentage Cumulative %
1–2 times 63 16.0 16.0
3–5 times 140 35.6 51.7
6–7 times 81 20.6 72.3
More than 7 times 109 27.7 100.0

4 OBJECTIVE 2: Service Quality Indicators

2.To determine the indicators for assessing the quality of transport services

4.1 Service Quality Index (SQI)

Table 8: Service Quality Index (SQI) Summary Statistics
Statistic Value
Mean SQI 2.37
SD 0.45
Median 2.40
Min 1.00
Max 3.20
% Excellent (≥3.5) 0.00
% Good (2.5-3.49) 41.98
% Fair (1.5-2.49) 52.67
% Poor (<1.5) 5.34

4.2 Quality Dimensions

Table 9: Service Quality Dimensions (Scale: 1=Poor, 4=Excellent)
Dimension Mean Score SD Rating
Reliability 2.46 0.68 Fair
Affordability 2.46 0.65 Fair
Courtesy 2.43 0.78 Fair
Safety 2.37 0.71 Fair
Comfort 2.15 0.74 Fair

4.2.1 Visualization: Quality Dimensions

4.3 Reliability Analysis (Cronbach’s Alpha)

Table 10: Internal Consistency Reliability (Cronbach’s Alpha)
Statistic Value
Cronbach’s Alpha 0.636
Raw Alpha 0.636
Standard Alpha 0.637
Average Correlation 0.260
Median Correlation 0.409
N Items 5.000

Notes: Cronbach’s Alpha >0.6 is considered acceptable for exploratory research. Inter-item correlations indicate moderate internal consistency among the 5 dimensions of the Service Quality Index.

4.4 Quality by Demographics

Table 11: Service Quality Index by Gender and Age Group
gender age_group n Mean SQI SD 95% CI Lower 95% CI Upper
Male Below 18 years 7 2.60 0.40 2.30 2.90
Female Below 18 years 7 2.54 0.56 2.13 2.96
Male 18–25 years 59 2.54 0.46 2.43 2.66
Male 36–45 years 39 2.46 0.36 2.35 2.58
Female 26–35 years 53 2.46 0.35 2.36 2.55
Male 26–35 years 47 2.43 0.44 2.31 2.56
Female 18–25 years 47 2.38 0.39 2.27 2.49
Male 56 years and above 7 2.29 0.47 1.93 2.64
Female 36–45 years 48 2.27 0.48 2.13 2.40
Female 46–55 years 35 2.20 0.46 2.05 2.35
Male 46–55 years 24 2.19 0.46 2.01 2.38
Female 56 years and above 20 2.02 0.56 1.78 2.26

4.4.1 Statistical Test: Gender Differences

Table 12: Independent Samples t-test - Service Quality by Gender
Group Mean.SQI t.statistic p.value Significance
Female 2.314 NA NA NA
Male 2.444 NA NA NA
Difference -0.129 -2.855 0.005 **

4.5 Correlation Matrix

5 OBJECTIVE 3: Transport Service Challenges

3.To identify the challenges involved in using the existing transport services

5.1 Problem Frequency

Table 13: Most Common Transport Service Problems (N=393)
Problem Count Percentage % Reporting Rank
Overcrowding 253 64.4 64.4% 1
Long Waiting Times 234 59.5 59.5% 2
High Fares 192 48.9 48.9% 3
Reckless Driving 149 37.9 37.9% 4
Poor Road Conditions 97 24.7 24.7% 5

5.1.1 Visualization: Problem Prevalence

5.2 Problems by Transport Mode

Table 14: Transport Problems by Primary Mode
primary_mode n Long Waiting (%) Overcrowding (%) High Fares (%) Reckless Driving (%) Poor Roads (%)
Trotro 367 59.1 67.3 48.2 38.4 25.3
Taxi 18 77.8 16.7 50.0 27.8 16.7
Tricycle 6 33.3 33.3 83.3 50.0 16.7
Motorcycle 2 50.0 50.0 50.0 0.0 0.0

5.3 Chi-Square Tests

Table 15: Significant Problem-Mode Associations (Chi-Square Tests)
Problem Mode Chi.Square df p.value Significant
X-squared1 Long Waiting Taxi 23.980 1 0 ***
X-squared2 Overcrowding Trotro 18.824 1 0 ***

5.4 Logistic Regression: Overcrowding

Table 16: Logistic Regression - Predictors of Overcrowding Experience
Predictor Odds Ratio 95% CI p-value Sig
age_group26–35 years 0.437 [0.21, 0.89] 0.024
monthly_incomeGHC 2,000 and above 0.363 [0.11, 1.08] 0.083
uses_trotro 9.049 [3.23, 28.7] 0.000 ***
congestion_levelModerate 14.006 [2.35, 124.27] 0.007 **
congestion_levelSevere 9.845 [1.8, 81.99] 0.015
congestion_levelVery severe 11.658 [2.04, 100.15] 0.011

Model Performance: - Pseudo R² (McFadden) = 0.11 - AIC = 459.1

5.5 Delay Frequency Analysis

Table 17: Frequency of Transport Delays in Reaching Destination
Delay Frequency Count Percentage Cumulative %
Rarely 27 6.9 6.9
Sometimes 163 41.5 48.3
Often 140 35.6 84.0
Always 63 16.0 100.0

6 OBJECTIVE 4: Mobility Challenges

  1. To understand the mobility challenges among users of the current transport services

6.1 Congestion Analysis

Table 18: Traffic Congestion Severity Perception (N=393)
Congestion Level Count Percentage
Severe 192 48.9
Very severe 114 29.0
Moderate 76 19.3
Low 11 2.8

6.1.1 Visualization: Congestion Levels

6.2 Daily Traffic Time

Table 19: Time Spent in Traffic Daily (One-Way Commute)
Daily Traffic Time (One-Way) Count Percentage
Less than 30 minutes 98 24.9
30–60 minutes 168 42.7
1–2 hours 95 24.2
More than 2 hours 32 8.1

6.3 Peak Traffic Times

Table 20: Most Common Peak Traffic Delay Periods
Peak Traffic Time Count Percentage
Evening [4–8 pm] 190 48.3
Morning [6–9 am] 132 33.6
Afternoon [12–3 pm] 70 17.8
Late night 1 0.3

6.4 Vulnerable Groups

Table 21: Groups Perceived to Face Greatest Mobility Challenges
Group Count Percentage Rank
Workers 269 68.4 1
Students 227 57.8 2
Traders/Market Women 188 47.8 3
Elderly 115 29.3 4
Persons with Disabilities 111 28.2 5

6.4.1 Visualization: Vulnerable Groups

6.5 Multiple Linear Regression: Travel Time

Table 22: Linear Regression - Predictors of Daily Travel Time
Predictor Beta Std. Error 95% CI t-value p-value Sig
age_group56 years and above 0.343 0.179 [-0.009, 0.696] 1.913 0.056
occupationStudent -0.277 0.130 [-0.532, -0.022] -2.135 0.033
congestion_numeric 0.413 0.052 [0.31, 0.515] 7.940 0.000 ***
study_routeEjisu– Kejetia Road 0.446 0.119 [0.212, 0.68] 3.751 0.000 ***
study_routeKejetia– Santasi Road -0.442 0.127 [-0.692, -0.191] -3.468 0.001 ***
uses_taxi 0.173 0.097 [-0.018, 0.365] 1.780 0.076
fare_per_trip 0.019 0.006 [0.007, 0.031] 3.201 0.001 **

Model Summary: - R² = 0.344 - Adjusted R² = 0.315 - F-statistic = 11.59 (p < 0.001)

6.6 Ordinal Logistic Regression: Service Quality

Table 23: Ordinal Logistic Regression - Predictors of Service Quality Rating
Predictor Odds Ratio 95% CI Lower 95% CI Upper p-value Sig
age_group26–35 years 1.916 0.993 3.695 0.052
age_group36–45 years 2.750 1.328 5.694 0.006 **
age_group46–55 years 2.035 0.915 4.525 0.081
education_levelTertiary (College/University/Polytechnic) 2.481 1.306 4.712 0.006 **
monthly_incomeGHC 1,000 – 1,499 0.246 0.099 0.613 0.003 **
monthly_incomeGHC 1,500 – 1,999 0.251 0.101 0.622 0.003 **
monthly_incomeGHC 2,000 and above 0.179 0.072 0.448 0.000 ***
monthly_incomeGHC 500 – 999 0.429 0.162 1.137 0.089
congestion_levelVery severe 0.142 0.034 0.589 0.007 **

Interpretation: Odds Ratios > 1 indicate increased likelihood of higher quality rating.

]

7 ADVANCED ANALYSIS: Machine Learning

7.1 Data Preparation

## ML Dataset prepared: 365 complete cases

7.2 Random Forest: Mode Switching

Table 24: Random Forest Performance - Mode Switching Prediction
Metric Value
Accuracy Accuracy 0.731
Sensitivity Sensitivity 0.723
Specificity Specificity 0.738
Precision Precision 0.680
F1 F1-Score 0.701

7.2.1 Feature Importance

7.3 K-Means Clustering: Passenger Segmentation

Table 25: Passenger Segment Profiles (K-Means Clustering)
cluster Size (n) % of Sample Avg Fare (GHS) Quality Index % Long Waiting % Overcrowding % High Fares % Trotro Users
1 120 30.5 7.0 2.5 27.5 100.0 24.2 99.2
2 116 29.5 7.7 2.5 41.4 0.0 33.6 84.5
3 102 26.0 8.1 1.9 96.1 87.3 85.3 96.1
4 55 14.0 23.1 2.6 100.0 80.0 67.3 94.5

7.3.1 Visualization: Cluster Profiles

7.4 Principal Component Analysis

Table 26: Principal Component Loadings (First 3 Components)
Variable PC1 PC2 PC3
afford_score -0.416 -0.556 0.119
reliability_score -0.474 -0.391 0.358
comfort_score -0.422 0.502 0.523
safety_score -0.455 0.529 -0.215
courtesy_score -0.466 -0.078 -0.733

Variance Explained: - PC1: 40.9% - PC2: 22.2% - PC3: 15.9% - Cumulative: 79%

8 DRIVER ANALYSIS

8.1 Driver Demographics & Operations

Table 27: Transport Operator Operational Characteristics
Metric Value
Total Drivers 339.0
Avg Experience (years) 6.1
Union Members (%) 66.4
Full-time Operators (%) 69.9
High Daily Trips (7+) (%) 54.3
Poor Road Conditions (%) 42.5
Frequent Congestion (%) 73.7
Encounters Authority Challenges (%) 19.8

8.2 Congestion Causes (Driver Perspective)

Table 28: Congestion Causes Identified by Drivers (N=339)
Cause Count Percentage Rank
Driver Indiscipline 270 79.6 1
Poor Traffic Control 265 78.2 2
Street Trading/Hawking 176 51.9 3
Poor Road Network 123 36.3 4
Lack of Parking Space 117 34.5 5

8.3 Text Mining: Authority Challenges

Table 29: Top 20 Words in Authority Challenge Descriptions
Word Frequency
extortion extortion 11
police police 7
always always 6
asking asking 6
drivers drivers 6
money money 6
stops stops 6
unwarranted unwarranted 6
bribe bribe 5
collection collection 5
arguments arguments 4
fines fines 4
harassment harassment 4
officers officers 4
sometimes sometimes 4
taking taking 4
unapproved unapproved 4
verbal verbal 4
collecting collecting 3
fees fees 3

8.3.1 Word Cloud

8.4 Improvement Suggestions

Table 31: Driver Suggestions for Mobility Improvement
Suggestion Count Percentage Rank
Better Traffic Management 299 88.2 1
Enforcement of Traffic Laws 291 85.8 2
More Parking/Loading Spaces 271 79.9 3
Construction of New Roads 115 33.9 4
Reduction of Street Hawking 101 29.8 5

9 COMPARATIVE ANALYSIS

9.1 Driver vs Passenger Perspectives

Table 32: Comparative Challenge Perspectives
Challenge Driver_Pct_Driver Driver_Pct_Passenger Passenger_Pct_Driver Passenger_Pct_Passenger
Poor Roads 36.3 0 0 24.7
Indiscipline 79.6 0 0 0.0
Street Trading 51.9 0 0 0.0
Poor Traffic Control 78.2 0 0 0.0
Long Waiting 0.0 0 0 59.5
Overcrowding 0.0 0 0 64.4
High Fares 0.0 0 0 48.9

9.1.1 Visualization: Perspective Comparison

9.2 Route Comparison

Table 33: Comparative Route Analysis
study_route Driver Sample % Always Congested % Poor Roads Passenger Sample Avg Fare (GHS) % Severe Congestion
Adum - Tafo Road 124 50.8 62.9 87 7.9 89.7
Ejisu - Kejetia Road 117 16.2 31.6 123 6.6 76.4
Kejetia - Santasi Road 96 55.2 28.1 76 8.8 61.8
Sofoline - Adum Road 2 0.0 100.0 107 15.6 81.3

10 POLICY RECOMMENDATIONS

10.1 Infrastructure Preferences

Table 34: Infrastructure Improvement Priorities
Improvement Support_Count Support (%) Priority
Expand Road Networks 269 68.4 High Priority
Improve Traffic Management 230 58.5 Medium Priority
Develop Bus Rapid Transit (BRT) 188 47.8 Medium Priority
Build Pedestrian Walkways 140 35.6 Lower Priority

10.2 Modern Transport Systems Support

Table 35: Public Support for Modern Transport Systems (Light Rail, BRT)
Response n Percentage
Do Not Support 57 14.5
Support 336 85.5

11 CONCLUSIONS

11.1 Key Findings Summary

11.1.1 Research Objective 1: Range of Transport Services

11.1.2 Research Objective 2: Service Quality Indicators

11.1.3 Research Objective 3: Service Challenges

  • Top 3 Problems:

11.1.4 Research Objective 4: Mobility Challenges

11.2 Machine Learning Insights

  • Random Forest Accuracy: for mode switching prediction